A user on the LocalLLM subreddit claims that some enthusiasts are already testing Gemma 5. This follows the current widespread use of models such as Gemma 4 and Qwen.